When cluster validation is executed on a cluster already configured, if tests are selected by default, will only be used to appraise the storage resources of disks that are offline or not assigned to an application or service cluster. This is based on a sheltered mechanism. The cluster validation wizard reports in cases where tests have been selected for storage but will not run into storage with online status, ie the storage used by clustered applications or services. And is designed to avoid disruption of applications or high-availability services that depend on these disk resources are online.
The test system configuration, inventory and testing network tests have very squat overhead and can be made without a perceptible effect on the servers in a cluster. The services of Microsoft Customer Support (CSS) may also request the execution of the validation of the cluster in a cluster of production as part of normal procedures for solving problems. In this scenario, the wizard will perform an inventory of hardware and software, conduct a network test and validate the system configuration. There may be some scenarios where you only need a subset of all tests. For example, if you are troubleshooting a network problem on a cluster of production, services of Microsoft Customer Support (CSS) may ask you to run only the inventory of hardware and software and network testing.
Another scenario in which the services of Microsoft Customer Support (CSS) may ask you to run the cluster validation in production is when an error cluster storage possibly by mistake or change in the underlying storage configuration. By default, the wizard warns you that have been selected storage tests but was not implemented in storage that is online, ie in storage used by applications or services in cluster. In this situation, you can perform a valid test to create or choose a logical unit number (LUN) of the new shared storage device and present it to all nodes. When testing this LUN, you can avoid the interruption of applications and services that are already clustered within the cluster online and even test the underlying storage subsystem.
If a failover cluster passed the inclusive set of validation tests and no hardware or software changes in future, then being measured a supported configuration. However, when making routine updates to software components, such as drivers and firmware, you may need to rerun the validation wizard to make certain that the current configuration supports failover cluster. The following instructions can help in this process:
- All components of the storage stack must be identical on all cluster nodes. These components consist of HBA, HBA firmware and drivers, software I / S and multi-path components of the device specific module (DSM).
- To minimize the impact on services and applications highly available, a recommended practice is to maintain a small LUN available to allow the Wizard to run validation tests on the available storage without adversely affecting the applications and services cluster. Thus, if the customer support services from Microsoft (CSS) you are asked to run a full set of cluster validation tests, the wizard will execute the default behavior and evidence in the available storage (only the new LUN).
Bookmarks